XYZ / April 4 Update

On March 31, the team met, and we experimented with the XY plotter.

This is what we discovered:

  • the XY area covered by the XY plotter is sufficient for our purposes. What is more critical is the size of the foam noodle we will be using. The XY plotter is 2 ft the foam noodle is 2ft – already that’s 4 feet radius of movement which exceeds our play area of 6ft by 6ft.
  • we hung the Makeblock Plotter with string by the ceiling and this arrangement insufficient because the weight of the plotter changes as the head moves. We will need to attach it directly on to the metal structure on the ceiling. We will email Rob to canvas for the space.
  • We determined the kind of part we will need to connect the Markblock (using the holes in the metal) to the ceiling’s metal hanging system. Tony will make a suitable block that will ensure a strong attachment.
  • We have determined it is important for the purposes of interaction that the robot have a district rotary and level-changing abilities. We flirted with the idea of just having a rotation because it would be easier, but we felt it would be compromising too much. Also, we would exclude children from the interaction if level changes were not possible. Shieya and I are coming up with ideas for a rotary/level-change system.

PDS / April Update

POWER

I initially waited to construct the new LED matrix until I had a grip on the power requirements. My old board was 2240 LEDs running from a 5V 70A 350W power supply, although it just good enough (I know this because when I run the whole thing at full white, it flickers). However, even as I type this I become unsure of my knowledge — Eric Rosenthal told me that when white LEDs turn yellow that means too much current. Overall, my deficiencies in understanding electricity is something that I wished to address, and measuring power requirements for the new board was a means of doing so.

I started out by scheduled a meeting with Eric Rosenthal, power guru of ITP. He recommended that I first find the “peak instantaneous  voltage” of my old LED board by connecting a 1 OHM resistor and using Ohm’s Law to calculate exactly what is the amperage that I require. The different patterns that I displayed drew different amounts of current, depending on how many LEDs were on and what colour the LEDS were.

However I didn’t really do a good job at measuring the current this way. The first 1OHM resistor I used was a tiny one, and when I made the circuit in series, nothing happened. I used other resistors and mostly, nothing happened on the board since there was not enough current. One resistor smoked and I suspect I blew other resistors also.

Later, after consulting with Jeff Feddersen, I learned that different resistors had different Wattage. I would need a much heartier resistor to do what I needed. However, even using 5W and 10W 1OHM resistors in the ITP shop, I still could not calculate what I needed. Time kept ticking and I just wanted to start making the damn thing.

So I decided that if my current power supply worked for a 2240-pixel LED matrix, it would definitely be enough to work on a 1850 and 1900-pixel LED matrix, since I decided to do it in two modules.

PROJECTION SURFACE 

I put up the metal L brackets according to plan and the paper was not as tight as I wanted. The thin part of the L bracket could be the culprit. Next steps on this will be to use nails with big heads to increase surface area that sandwiches the paper.

POWER ON THE SIDE

From previous experience I know that each LED strip needs its own connection to the power supply, rather than running the power through the entire strip (which does not do the job). However when I was using a stiff piece of plywood to mount the LED strips to, I could just staple the 12 AWG wire to the board and everything wouldn’t move, including the solder points.

But since I am using a flexible piece of vinyl, I face the challenge of needing a stiff 12 AWG wire soldered to each LED strip, however it needs to be attached to the flexible vinyl. Something stiff attached to something flexible. I consulted Ben Light on this. After the consultation and two hours of deep visualization while sitting on a display rack in Home Depot, I decided upon a low tech solution: I would staple a piece of plywood to the back of the vinyl with heavy-duty staples. And I would staple the 12 AWG to it too. Just like on the old board.

Future Scenarios / Executive Summary of Research

EXECUTIVE SUMMARY — HOW SOCIAL MEDIA, AI & IOT AFFECT THE TRAJECTORY OF POLITICAL SYSTEMS AND IDEOLOGIES

“If you want to change politics, you first have to change culture, because politics flows from culture.”

-Christopher Wilke, former data scientist at Cambridge Analytica

Like all things, political systems and ideologies are marked by a constant state of change. The nature of these changes depend upon the variety of factors that come to bear on a specific historical situation: historical events, technological change, philosophical ideas, religious ideas, human migration patterns, diseases, exchanges between different cultures… etc. Each of these factors has a role in shaping peoples’ ideology, which can be defined as “a set of beliefs about the proper order of society and how it can be achieved.” The collective ideology of a given population is known as an ideological landscape.

This executive summary argues that the leap in communication technologies marked by the advent of social media platforms (Facebook, Twitter, Youtube, Instagram) represent:  a game-changing shift in how political systems and populations interact with each other, even if the content of that interaction remains based on historical and social factors. This summary argues that the complexity of this shift can be seen in the variety of ways that social media has been used by different political actors from the time period 2011-2017 to directly influence the trajectories of political systems, the ideological landscape they rest upon, as well as the units of that landscape — individual people.

As such, the first two case studies cited by my research present two opposing scenarios: one in which social media is used by a democratic movement to influence an authoritarian political system (2011 Egyptian Revolution); and another where social media is used by an authoritarian political system to influence the population of a democratic country (Russian interference in US politics, 2014 – present). Further improvement in machine learning has aided efforts by political actors to use social media to influence populations. This is seen in our third case study (Cambridge Analytica).

As we look toward the future, based on evidence from the 2011-2017 time period, it is reasonable to that data collection and analysis capabilities from machine learning will continue to be utilized by political actors. As such, the potential for the Internet of Things (IOT) in the context of the power of mass collection of data, will have further disruptive effects on politics.

Case Study 1: 2011 Egyptian Revolution

Egypt was among the first prominent examples of social media being central to a large-scale political upheaval. Indeed, the 2011 Egyptian Revolution was dubbed the “Facebook Revolution”. The Pew Research Center argues that social media platforms facilitated not just the Egyptian revolution, but the entire Arab Spring. It did so in three main ways:

Networks formed online were crucial in organizing a core group of activists, specifically in Egypt.

Civil society leaders in Arab countries emphasized the role of “the internet, mobile phones, and social media” in the protests.

Additionally, digital media has been used by Arabs to exercise freedom of speech and as a space for civic engagement.”

Indeed, social media shaped the nature of the protests themselves to make them more powerful – yet more unpredictable. Scholar Essam Mansour writes, “What makes social media such a powerful and unpredictable force in global politics is that they replace the need for a charismatic leader. Certainly, there were a number of outstanding and courageous protesters, but there was no single face attached to this revolution; They are impossible to control or shut down.” In a way, social media enabled the protests to behave like a swarm or flock, where all the units behave individually, yet somehow end up acting collectively.

However, Mansour also describes social media as “politically agnostic”, meaning that “there is no overseer that watches over, or brand that is stamped on a grass-roots social network. The best anyone can hope for is a small say in what goes on within it.”

Case Study 2: Russian Interference in US Politics

In May 2017, special counsel Robert Mueller was appointed to investigate possible collusion between Russia and the election campaign of US President Donald Trump. The investigation concluded that there was in a strategic effort made by Russian operatives to sow discord in the U.S. political system through using social media to disseminate fake news from as far back as 2014. Russian agents posed as US citizens, made social media groups, organized protests and created web sites with misinformation designed to exploit and exacerbate polarization amongst US citizens. Over time, these social media accounts reached significant numbers of Americans and helped the Trump Campaign defeat Hillary Clinton.

That said, it is unknown how much the effort helped the Trump campaign when compared to other factors. Some argue that the Russian troll farm wasn’t actually that effective or insidious, and they were doing very standard practices of audience development that any media company would be doing.

Cast Study 3: Steve Bannon, Cambridge Analytica & Culture War

Cambridge Analytica (CA) is the intersection of three things: big data, behavioural psychology and micro-targeting. “Big Data” is “an aggregation of all data points that you can get your hands on.” Micro-targeting refers to individualized messaging. So, instead of classifying a person as belonging to a particular voter class (i.e. latino voters or women voters), as has been done in traditional political analysis, micro-targeting allows for a specific message to be tailored to a specific personality. CA is credited with having assisted the Ted Cruz campaign, and then the Trump campaign in 2016.

In an interview with The Guardian, former CA data scientist Christopher Wilke discussed the core issue on how social media, data, and AI can affect not just political system but the ideological landscape itself. He described how Steve Bannon helped to start CA to be a “psychmetric weapon” to fight a culture war in the United States. Bannon was the former editor of Breitbart News, an organization that holds views associated with the Alt-Right that is highly critical of mainstream media for being unfairly liberally biased.

Wilke stated that Steve Bannon “the reason he was interested in this [psychometric weapon], is he had this idea of the Brietbart doctrine, which is that if you want to change politics, you first have to change culture, because politics flows from culture. And so what i said is that if you want to change culture, you first have to understand the units of culture. People are the units of culture. So if you want to change politics, you first have to change people, to change the culture.” 

Trump’s success in the 2016 election as well as the success of the Breitbart new organization in general, can be seen as evidence of the success of this effort at changing the changing the individual “units of culture” using data, AI and social media.

Thus, the ideological landscape of the US has been, and is currently being, shifted by the interaction of social media (interface to the people), data (gathering of information on the people) and AI (analysis of information about the people).

CONCLUSION

In reviewing the case studies from above, it appears the description of social media as “politically agnostic” applies to big data and AI as well. The above evidence supports the interpretation of social media, big data and AI as representing game-changing increases in capabilities of political actors. As we have seen, for the political actors who have effectively utilized these technologies — they are given a significant advantage over their competition.


Changing people to change culture is not a new concept. In the 20 century, communist governments sough schemes to re-educate their populace to fit their political system. (China’s Cultural Revolution is one prime example of such efforts). As such, based on the examples above it is difficult to say how these technologies impact “ideology” in any one-sided way. This is contrary to earlier arguments of how new technologies were inherently “democratizing”.

The only thing that can be concluded is that new technology seems to be the primary means of political struggle in the foreseeable future, as in the example cites above, those who are not ahead are left behind.

XYZ final project idea

 

For our final, Shreiya, Chengtao and I have decided to build a “dodge stick” game, where the objective of the game is to avoid a slow moving but relentless stick aimed at touching you. The inspiration for this came from the movement practice of Ido Portal, who famously trained UFC fighter Conor McGregor.

The X and Y of contraption will be approx. 6×6 feet in width and length. The part of it that interacts with the player will be a foam noodle that extends down from a cardboard Z that will be moved up and down by a threaded rod inside. For the safety of the player, we will ensure that no metal parts will come into contact with them. As for control, we are envisioning some kind of controller device that a Player 2 can use to guide the stick and try to touch the Player 1 who is inside the game.

See rough sketch below:

Issues to be resolved include:

-the stick portion will need to change levels AS WELL AS rotate. So it will be XYZR (R for rotation, and Z for level change). How will we make a system that allows for rotation and level changes?

-what material will we use for the gantry? None of us have much money — we are not going to shell our hundreds of $$$ for an XYZ final. How do we achieve our ends in an affordable way?

 

Solar Project: One week of stats!

On Tuesday March 13, we met up to assemble and install our solar project on the east facing windows on the ITP floor.

There was some concern that the circuit wouldn’t work, because when we connected the circuit (the Adafruit solar panels, the power converter, and the MRK1000), it never really worked even with the direct indirect sunlight (i.e. sunlight reflected from windows of the nearby building). However, we assembled the circuit at the north-facing windows. See the spotty reception below:

We tried using some capacitors to mediate the charge build-up but to no avail. Finally, it was decided that since it had worked when we had shown the project during the class, that it would work again with direct (not indirect) sunlight. Thus, we found a suitable cardboard enclosure to put it all in, and secured it to the window.

To our satisfaction, the device started sending data to the server on MyDevices which Dan had made an account for. It has been tracking the data since we put it in.

http://128.122.6.159/

On the data, it means:

Channel 0 means time
Channel 1 stats the battery level (which we am not sure we measured properly , because we are not using those pins to connect our solar panel.)
Channel 2 is sending 1 everything he can

Solar Energy Project (so far)

For our solar energy project, Marcha, Daniel & I are creating a solar-powered “status of Roland” program. What is that, you may ask?

Well quite simply, when we were gathering for our meeting last Saturday on the floor, Roland happened to be sitting with us and he basically offered unsolicited feedback on our ideas, to the point of discouraging hilarity. In other words, he shat on all our ideas (pardon my French).

In honour of this hilarious event, decided our program to operate like this:

  • when the solar panels reach the required power levels, it powers the MKR1000
  • Over wifi, the MKR1000 program communicates with a serve that displays an image of Roland with his thumbs up
  • when the solar panels do not reach the required power levels, the program will have Roland with his thumbs down

We will be using the shop’s Adafruit Medium 6V 2W solar panel, with its Li-Po charger.

What has been done:

  • We decided to not use a battery. Instead, we are relying on the power on/off of the board to run the program. This simplifies the circuit.
  • MKR1000 power measurements by multimeter:
    • without wifi: 20mA
    • with wifi: 120mA
    • (above calculations to be verified)
  • Solar panel when directly measured with multimeter (i.e. not through the Li-Po charger) is 5V but is unable to light an LED.

A Tangent that I Spent A Long Time On

  • I spent a lot of time on the question “how long can a 9V 200 mAh battery power a 3.3V 200mA board?” I couldn’t compute this.

Problem to be overcome:

  • Due to a combination of scheduling problems (most of us are available at times where we don’t have much light) as well as weather patterns (cloudy days, snowy days) we have not been able to experiment with the solar panel during times of ample daylight and thus our experimentation with the solar panel has been limited.
  • As a result we have not been able to get the MKR1000 running using just the sun.
  • We have not even been able to get an accurate reading of the solar panels Voltage or Current with direct sunlight
  • Daniel has set up the server through his knowledge of servers through his Networked Interactions class (thanks Daniel)

Next Steps

  • Next steps is:
    • come to ITP bright and early and take reading of solar panel with direct sunlight
    • build circuit to supply necessary voltage to MKR1000 without frying it (may have to use voltage regulator, given it’s operating capacity is 3.3V)
    • come in bright and early again and test the circuit in real time with real sunlight.